MEDB 5510, Module 10, Data collection

Topics to be covered

  • What you will learn
    • Focus groups, interviews
    • Questionnaires
    • Likert scales
    • Nominal scales
    • Case report forms
    • Advantages of different data collection methods

10-01, Running focus groups

  • Two person job
  • Facilitator
    • Asks questions
    • Guide discussion, but stays neutral
  • Recorder
    • Runs the tape recorder
    • Keeps written notes
  • Debriefing session

10-01, Structure of a focus group

  • Six to ten participants
  • Multiple groups
    • Stop when you achieve saturation
  • Sixty to ninety minutes
  • Ten questions, semi-structured interview

10-01, Interview

  • Administration
    • Telephone
    • Face-to-face
  • Format
    • Unstructured
    • Semi-structured
    • Structured

10-01, Example of structured and unstructured interviews

Break #1

  • What you have learned
    • Focus groups, interviews
  • What’s coming next
    • Questionnaires

10-02, Questionnaire

  • Not the same as a survey
    • Questionnaire is the form
    • Survey is the process
  • Administration
    • Mail
    • Computer (REDCap)
    • Telephone
    • In person

10-02, Question types

  • No one to supervise if person answers “incorrectly”.
    • Q. What is your age? A. Older than dirt.
    • Q. What is the one word that describes your greatest weakness? A. Inability to follow directions.
  • Three broad classes of questions
    • Open ended
    • Partial open ended
    • Close ended

10-02, Advantages of open ended questions

  • Richer, more detailed responses
  • Captures unanticipated responses
  • Avoids “pigeonholing”.
    • “Pigeonholing is a process that attempts to classify disparate entities into a limited number of categories (usually, mutually exclusive ones). The term usually carries connotations of criticism, implying that the classification scheme referred to inadequately reflects the entities being sorted, or that it is based on stereotypes.” Wikipedia.

10-02, Advantages of close ended questions

  • Faster
  • Easier to analyze
  • Better quality control

Note: partially open ended questions are a compromise and have the worst of both worlds.

10-02, Types of close ended questions

  • Fill in the blank
  • Ordinal scales
    • Semantic differential scales
    • Likert scale
    • Visual analog scale
  • Nominal scales
    • Select one option
    • Select all that apply

10-02, Semantic differential scale

  • Pairs of polar opposite adjectives
  • Internal scale with no intermediate descriptors
  • Select a response along the continuum.
  • Common themes
    • Evaluative (good-bad)
    • Potency (strong-weak)
    • Activity (fast-slow)

10-02, Example of a semantic differential scale

Semantic differential scale with three items

Break #2

  • What you have learned
    • Questionnaires
  • What’s coming next
    • Likert scales

10-03, Visual analog scale

  • Line segment
    • Total length 100 millimeters
  • Labels at two extremes
    • Alert/Extremely drowsy
    • Energetic/worn out
    • completely full/starving
  • Mark an X along the continuum

10-03, Example of a visual analog scale

Visual analog scale for pain

10-03, Likert scale

  • Discrete values
  • Extremes and intermediate points labelled.
  • Measurements of frequency
    • Very frequently, Frequently, Occasionally, Rarely
    • Never, Rarely, Sometimes, Often
    • Every day, Two or three times a week, Once a week, Less than once a week

10-03, Likert scale

  • Satisfaction (Very satisfied, satisfied, dissatisfied, very dissatisfied)
  • Quality (Very good, Good, Fair, Poor)
  • Agreement scale (5 points)
    • Strongly disagree, disagree, neutral, agree, strongly agree
  • Agreement scale (6 points)
    • Strongly disagree, disagree, slightly disagree, slightly agree, agree, strongly agree

10-03, Example of a Likert scale

10-03, Likert response options

  • Number of points
    • Larger is better? worse? than smaller
    • Even is better? worse? than odd
  • Not applicable, Don’t know
  • Reverse scaling

Break #3

  • What you have learned
    • Likert scales
  • What’s coming next
    • Nominal scales

10-04, Nominal responses

  • Pick one of the following
    • Assumes mutually exclusive
  • Pick all that apply
    • Allows for overlap
    • Code as multiple binaries
  • Code for “other”
    • Partial open ended
  • Code for “Prefer not to answer”

10-04, General structure of a questionnaire

  • Sensitive questions last
  • Broad categories
    • Income
  • Do you really need to ask?
  • Avoid leading up to a question

10-04, Practical advice

  • Don’t ask for what you don’t need
    • More data than you have time to analyze
  • Make your questionnaire easy for data entry
  • Pilot test your questionnaire

Break #4

  • What you have learned
    • Nominal scales
  • What’s coming next
    • Case report forms

10-05, Case report form (CRF)

  • Paper or electronic form
    • Prospective studies
    • Retrospective chart reviews
  • Used to collect and store data
    • Filled out by researcher
    • One form per patient

10-05, What sort of data

  • Medical history
  • Physical
  • Potentially other types of data
  • No personal identifiers

10-05, Design issues

  • Clear
  • Unambiguous
  • Easy to read
  • Consistent format

Break #5

  • What you have learned
    • Case report forms
  • What’s coming next
    • Advantages of different data collection methods

10-06, Existing data sources

  • Re-analysis of published data
    • Protect confidentiality
    • Give appropriate acknowledgement or co-authorship
  • Curated databases
    • Contractual limitations
    • Sample weights
  • Web scraping
    • Respect limits on volume and speed of download
    • Protect copyrights

10-06, Advantages of focus groups

  • Faster than separate interviews
  • Participants build each others comments

10-06, Advantages of interviews

  • On-the-spot fixes
  • Increased participation rate

10-06, Advantages of a questionnaire

  • Less labor
  • No interviewer bias

10-06, Advantages of existing data

  • Fast
  • Cheap
  • Naturalistic

Summary

  • What you have learned
    • Focus groups, interviews
    • Questionnaires
    • Likert scales
    • Nominal scales
    • Case report forms
    • Advantages of different data collection methods